率的非劣效试验中样本含量的估计——SAS和PASS实现

目的 提供二分类定性资料平行设计非劣效临床试验样本含量最常用的计算公式及其 SAS和PASS过程，并为相关参数的设置提供参考。方法 基于二项分布的正态近似理论推导样本含量的估计公式，通过SAS程序和PASS过程探讨各重要参数（样本率、非劣效界值）变化时样本含量及检验效能的变化情况。结果 对率的非劣效试验样本含量的计算，公式、SAS程序和PASS过程能得到一致结果；当检验水准和对照组样本率确定时，试验组样本率越大、检验效能越小、界值越大，所需样本含量越小。结论 利用本文提供的公式、SAS程序和PASS过程，可以帮助研究者系统快速得到二分类资料2组平行非劣效设计时的样本含量。试验组样本率、检验效能和非劣效界值是非劣效临床试验估计样本含量必须认真考虑的参数。

Sample size calculation of dichotomous variables in non-inferiority clinical trials-SAS and PASS realization

Objective The aim of this study was to provide the most commonly used formula on sample size in rate non-inferiority clinical trials and its SAS and PASS realization procedure, in order to provide suggestion of related parameters in calculation of sample size. Methods The formula as well as present SAS and PASS procedures were derived for sample size calculation based on the normal approximation theory of binomial distribution. And the change of sample size was explored with the change of important parameters, i.e., sample rate, power and non-inferiority margin. Results The result of the formula was consistent with that of SAS program and PASS process. When the significance level and the control group rate were determined, the larger the test group rate, the smaller the power of test, the larger the margin value, the smaller the required sample size. Conclusion The formula, SAS procedures and PASS process illustrated in this paper can help researchers obtain the sample size of dichotomous variables in non-inferiority clinical trials systematically and conveniently. The rate of test group, the power and the non-inferiority margin must be carefully considered in non-inferiority clinical trials.